106 ALASKAN GLACIER STUDIES and partly because of its relation to Variegated Glacier, the eastern margin of the Hub-bard Glacier was studied in some detail in 1905; and in view of the fact that notable changes had taken place when we studied it again in 1909, its characteristics will be stated here in some detail. Within the area of observation this margin may be divided into three portions, an upper, middle and lower portion. In the upper portion, which lies within and just outside of the mountain valley, there is a border of lateral moraine of considerable breadth, which near the very margin is so thick that most crevasses are filled, making it possible for one to travel over the surface. But a short distance out on the ice the surface becomes badly crevassed and the moraine much thinner. Between the glacier and the mountain there is a marginal valley in which a small stream flows, and this is continued westward by a low, undulating depression in which lie several marginal lakes (PI. XLVII, B.) This depression is an interlobate area between the Hubbard and Variegated Glaciers, and its outer portion, where the two glaciers coalesce, is underlaid by ice. Below the depression, for a distance of about two miles, the Hubbard and Variegated Glaciers are united and one cannot state exactly where the line between the two should be drawn (PI. XLII). Nevertheless it is perfectly plain that all but a very small area along the exact boundary is to be assigned to one or the other of the glaciers. This is made evident by the character and distribution of the moraines. Hubbard and Variegated Glaciers approach each other at very nearly right angles, but Variegated Glacier spreads out in a broad, stagnant bulb, covered with moraine from side to side, while Hubbard Glacier pursues its course to the sea as a vigorous, actively-moving glacier, uncrevassed and moraine-covered only on the very margin. The marginal morainic ribbons on the Hubbard bend as if the. Variegated dominated; but, in view of the stagnant state of the Variegated, this cannot be the case and we must assume that the Hubbard is set in a valley of sufficient depth to deflect it westward and prevent it from invading the area occupied by the stagnant Variegated, toward which it is directed when it comes from its mountain valley. On the two sides of the area of coalescence the moraines are wholly different. On the Hubbard side there are bands parallel to the direction of ice motion toward the fiord and in these morainic bands fragments of red gneiss are sufficiently predominant to give the moraine color; but on the Variegated side the ribbon appearance is absent and in its place there is a confused waste of ablation moraine blackish in color because of the great abundance of black hornblende gneiss fragments, and with a roughly-crescentic banding apparent when viewed from a distance. There is a noteworthy difference in glacier condition on the two sides of the area of junction of the Variegated and Hubbard Glaciers. Where the two unite the glacier surface is so deeply covered with moraine that the ice is in large part hidden, though it appears here and there in crevasses and in cliffs. The emergence of springs, the abundant evidence of slumping of the moraine, and the general absence of vegetation all testify to the presence of ice beneath those areas in which no ice is visible. Usually we could also prove the existence of ice in such places by thrusting our ice axes down to it through the morainic veneer (PI. XLVIE, A). The irregular protection which the moraine cover offers to the underlying ice here, as elsewhere, gives rise to the development of hillocks and ridges, with intervening kettles and valleys. This condition of the glacier extends not only over the junction of the two glaciers, but also southward over the stagnant outer bulb of the Variegated Glacier.